TY - JOUR

T1 - The admixed population structure in Danish Jersey dairy cattle challenges accurate genomic predictions

AU - Thomasen, Jørn Rind

AU - Sørensen, Anders Christian

AU - Su, Guosheng

AU - Madsen, Per

AU - Lund, Mogens Sandø

AU - Guldbrandtsen, Bernt

PY - 2013/7

Y1 - 2013/7

N2 - The main purpose of this study is to evaluate whether the population structure in Danish Jersey known from the history of the breed also is reflected in the markers. This is done by comparing the linkage disequilibrium and persistence of phase for subgroups of Jersey animals with high proportions of Danish or US origin. Furthermore, it is investigated whether a model explicitly incorporating breed origin of animals, inferred either through the known pedigree or from SNP marker data, leads to improved genomic predictions compared to a model ignoring breed origin. The study of the population structure incorporated 1,730 genotyped Jersey animals. In total 39,542 SNP markers were included in the analysis. The 1,079 genotyped bulls with de-regressed proof for udder health were used in the analysis for the predictions of the genomic breeding values. A range of random regressions models that included the breed origin were analyzed and compared to a basic genomic model that assumes a homogeneous breed structure. The main finding in this study is that the importation of germ plasma from the US Jersey population is readily reflected in the genomes of modern Danish Jersey animals. Firstly, linkage disequilibrium in the group of admixed DJ animals is lower compared to the groups of the original DNK and USJ animals. Secondly, persistence of linkage disequilibrium phase is not conserved for longer marker distances between animals with mainly Danish or US origin. The overall conclusion from comparing different genomic prediction models showed that including population structure in a random regression prediction model, did not clearly improve the reliabilities of the genomic predictions compared to a basic genomic model

AB - The main purpose of this study is to evaluate whether the population structure in Danish Jersey known from the history of the breed also is reflected in the markers. This is done by comparing the linkage disequilibrium and persistence of phase for subgroups of Jersey animals with high proportions of Danish or US origin. Furthermore, it is investigated whether a model explicitly incorporating breed origin of animals, inferred either through the known pedigree or from SNP marker data, leads to improved genomic predictions compared to a model ignoring breed origin. The study of the population structure incorporated 1,730 genotyped Jersey animals. In total 39,542 SNP markers were included in the analysis. The 1,079 genotyped bulls with de-regressed proof for udder health were used in the analysis for the predictions of the genomic breeding values. A range of random regressions models that included the breed origin were analyzed and compared to a basic genomic model that assumes a homogeneous breed structure. The main finding in this study is that the importation of germ plasma from the US Jersey population is readily reflected in the genomes of modern Danish Jersey animals. Firstly, linkage disequilibrium in the group of admixed DJ animals is lower compared to the groups of the original DNK and USJ animals. Secondly, persistence of linkage disequilibrium phase is not conserved for longer marker distances between animals with mainly Danish or US origin. The overall conclusion from comparing different genomic prediction models showed that including population structure in a random regression prediction model, did not clearly improve the reliabilities of the genomic predictions compared to a basic genomic model

KW - cattle

KW - genomic prediction

KW - population structure

U2 - 10.2527/jas.2012-5490

DO - 10.2527/jas.2012-5490

M3 - Journal article

C2 - 23658363

VL - 91

SP - 3105

EP - 3112

JO - Journal of Animal Science

JF - Journal of Animal Science

SN - 0021-8812

IS - 7

ER -